Letter pubs.acs.org/NanoLett
Nanoporous Designer Solids with Huge Lattice Constant Gradients: Multiheteroepitaxy of Metal−Organic Frameworks Zhengbang Wang,† Jinxuan Liu,† Binit Lukose,‡ Zhigang Gu,† Peter G. Weidler,† Hartmut Gliemann,† Thomas Heine,‡ and Christof Wöll*,† †
Karlsruhe Institute of Technology, Institute of Functional Interfaces (IFG), Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany ‡ Center for Functional Nanomaterials, School of Engineering and Science, Jacobs University Bremen, Campus Ring 1, 28759 Bremen, Germany S Supporting Information *
ABSTRACT: We demonstrate the realization of hierarchically organized MOF (metal−organic framework) multilayer systems with pronounced differences in the size of the nanoscale pores. Unusually large values for the lattice constant mismatch at the MOF−MOF heterojunctions are made possible by a particular liquid-phase epitaxy process. The multiheteroepitaxy is demonstrated for the isoreticular SURMOF-2 series [Liu et al. Sci. Rep. 2012, 2, 921] by fabricating trilayer systems with lattice constants of 1.12, 1.34, and 1.55 nm. Despite these large (20%) lattice mismatches, highly crystalline, oriented multilayers were obtained. A thorough theoretical analysis of the MOF-on-MOF heterojunction structure and energetics allows us to identify the two main reasons for this unexpected tolerance of large lattice mismatch: the healing of vacancies with acetate groups and the low elastic constant of MOF materials. KEYWORDS: Metal−organic frameworks, heteroepitaxy, lattice mismatch, thin films, loading ighly porous, crystalline “Designer Solids” with different functionalities fabricated via the assembly of molecular connectors and linkers are receiving a still increasing amount of interest. Going beyond the adsorption of small molecules for gas storage1−3 and separation,4 more advanced applications exploit the enormous flexibility of these metal−organic frameworks, MOFs, for applications in sensorics,5,6 optics,7,8 and energy storage.9−11 The countless number of different MOF types resulting from the combination of metal or metal− oxo connectors with organic linkers can be extended further by employing heteroepitaxy. Growing different types of MOFs in a sequential fashion12−15 allows entering the next level of complexity. Of particular interest here is the stacking of MOFs with different lattice constants. For example in sensorics, it may be required to have a first layer of MOFs with pore sizes sufficiently large to accommodate nanoobjects such as proteins16 or metal−nanoparticles. An additional topping layer with smaller pore sizes could then serve as a filter to reduce cross-talk in MOF-based sensorics.17 The range of lattice constants so far realized in MOFs ranges from less than 1 to 10 nm.16 A combination of MOF materials with different lattice constants is a major challenge, because heteroepitaxy of conventional materials requires lattice constant matching, that is, the lattice constant of the material to be grown must match that of the templating substrate. Particularly important is the case of metal-on-semiconductor epitaxy, where sophisticated strategies are used to overcome the restrictions
H
© 2014 American Chemical Society
imposed by the fact that the second material must match the lattice constant of the first material by better than 2%.18 When the lattice constants differ, the deposited layers are strained, leading to a termination of the growth or low quality of the epitaxially grown layers resulting from high defect densities. Here, we demonstrate that the high elasticity of MOFs strongly relaxes, if not removes, the requirement of lattice constant matching. We show that epitaxial growth can be obtained for lattice constant mismatches as large as 20%. Quantum chemical calculations allow determining the related interface energy penalty. The theoretical results also demonstrate that the intrinsic elasticity of MOFs helps to yield epitaxial growth also in case of strongly different lattice constants. For the heteroepitaxial growth of the hierarchically organized MOF multilayer systems, we use the liquid phase epitaxy (LPE) process, a layer-by-layer procedure that yields oriented, highly crystalline MOF thin films that are referred as SURMOFs.19 Briefly, the SURMOF process consists of spraying appropriately functionalized substrates with solutions of the reactants in a sequential fashion (see Figure S1, Supporting Information). This step-by-step process allows us to separately control the metal and organic linker deposition on the substrate. A huge Received: December 23, 2013 Revised: February 4, 2014 Published: February 10, 2014 1526
dx.doi.org/10.1021/nl404767k | Nano Lett. 2014, 14, 1526−1529
Nano Letters
Letter
advantage of this layer-by-layer process is the possibility to change the composition of the SURMOFs along the vertical direction. In previous work, we have demonstrated the heteroepitaxy of layer-pillar MOFs exhibiting the same lattice constant but different metal ions in the connectors.15 In order to investigate the importance of lattice mismatch in MOF-onMOF heteroepitaxy, here we keep the same connector but used different organic linkers, resulting in changing the lattice constant of the MOF thin films for yielding hierarchically organized MOF multilayer systems. The studies to be presented in this communication have been carried out using the structural platform provided by SURMOF-2, an isoreticular class of metal organic frameworks described in previous work.20 SURMOF-2 is synthesized from Cu-acetate and dicarboxylic acids using liquid phase epitaxy;20 the crystal structures of these compounds are shown in Figure S2 in the Supporting Information. In a first set of experiments, an initial MOF-layer has been grown by LPE from BPDC (biphenyl-4,4′-dicarboxylic acid) and Cu-acetate using the conditions that were published elsewhere.21,22 As evidenced by the out-of-plane (Figure 1, green curve) and in-plane (Supporting Information Figure S3, green curve) X-ray diffractograms, this process yields a crystalline, highly oriented SURMOF-2 with a lattice-constant of 1.55 nm. The out-ofplane XRD data reveals the crystallinity perpendicular to the substrate surface. The fact that only the (001)C and (002)C diffraction peaks are seen, demonstrates the orientated nature of the SURMOF. After 10 deposition cycles, the long BPDC linker was replaced by the shorter NDC (naphthalene-2,6dicarboxylic acid) linker. No change was made to the connector. As evidenced by the X-ray diffractograms (peaks labeled (001)B and (002)B) shown in Figure 1, after another 10 deposition cycles with the shorter NDC ligand a second MOFlayer with a smaller lattice constant of 1.34 nm was formed on the first one. A comparison to the XRD-data recorded for the corresponding SURMOF made directly using Cu-acetate and NDC reveals that the structures are identical. Note, the XRDpeaks located at 5.9° demonstrate that the initial SURMOF layer is still present. In the next step, an even shorter ligand, BDC (benzene-1,4-dicarboxylic acid), was used. Again, the XRD data (peaks labeled (001)A and (002)A) shown in Figure 1 (black curve, out-of plane) and Supporting Information Figure S3 (black curve, in-plane) demonstrate that a third type of MOFs with the same orientation but with a lattice-constant decrease to 1.12 nm has been grown on the supporting substrate (Cu-NDC). The well-defined multilayer structure of the trilayer is directly demonstrated by the SEM micrographs shown in Figure 2. Whereas the cross sections of the hierarchically porous MOF crystal films (Cu-BPDC + Cu-NDC+ Cu-BDC) reveal only little contrast (Figure 2A), selective staining23,24 can be used to make the different porosities of the adlayers visible. To this end, the sample shown in Figure 2A was immersed in a solution of a metal−organic compound, Eu(bzac)3bipy (bzac = 1-benzoylacetone, bipy = 2,2′-bipyridine, for details of staining process see Supporting Information) with a diameter of 0.97 nm. Because of its fairly large size, the Eu-compound is loaded only into the largest pores (Supporting Information Figure S6), and the lower Cu-BPDC-layer (thickness ∼250 nm) can be clearly identified (Figure 2B). The very effective spray-process22 used here yields a thickness of ∼250 nm with only 10 cycles. In order to demonstrate that the constraint of lattice matching is virtually nonexistent for this reticular class of
Figure 1. (Left) Schematic illustration of the construction of hierarchically porous MOF crystal films (Cu-BPDC + Cu-NDC + Cu-BDC) with large lattice mismatches grown on self-assembled monolayer (SAM) modified Au-substrates by using layer-by-layer LPE method (dark blue circles represent metal connectors). Upper right: Molecular structures of the organic ligands. Lower right: Out-of-plane XRD data recorded for the SURMOFs depicted on the left.
MOFs, we have instead of stepwise decreasing the lattice constant as in the example presented above also realized the opposite scenario of increasing lattice constants. Starting with an initial layer of Cu-BDC (with a pore-size of 1.12 nm) we continued to deposit Cu-NDC and finally Cu-BPDC, thus increasing the pore-sizes to 1.34 nm (NDC) and finally 1.55 nm (BPDC). The resulting scenario is shown in Supporting Information Figure S4. As confirmed by the out-of-plane (Figure S4, Supporting Information) and in-plane (Figure S5, Supporting Information) X-ray diffractograms, also for this case of increasing lattice constants the heteroepitaxial growth 1527
dx.doi.org/10.1021/nl404767k | Nano Lett. 2014, 14, 1526−1529
Nano Letters
Letter
Figure 2. (top) Scheme of nanoparticles size-selected loading in hierarchically porous SURMOFs. Down: (A) SEM image recorded for a hierarchically porous SURMOF (Cu-BPDC + Cu-NDC + Cu-BDC) with 5 nm gold films coated on the surface. (B) SEM image recorded for a hierarchically porous SURMOF (Cu-BPDC + Cu-NDC + CuBDC) after loading Eu(bzac)3bipy (bzac = 1-benzoylacetone, bipy = 2,2′-bipyridine) compound without coating gold films on the surface.
Figure 3. (top) Schematic illustration of the formation of flexible lattice matching junction in MOF-on-MOF epitaxy. (bottom) Molecular arrangement as obtained from the quantum chemical calculations (see text) for the heterojunction between Cu-BDC and Cu-NDC.
proceeds without difficulties despite the rather pronounced lattice mismatch. To better understand why for the MOF-on-MOF heteroepitaxy the lattice matching requirement appears to be less stringent than for other cases of heteroepitaxy, we have carried out an extensive set of quantum chemical calculations for the heterojunction between Cu-BDC and Cu-NDC. The method used here is an approximate variant of density functional theory (DFT), London dispersion-corrected self-consistent charge density-functional based tight-binding (DFTB).25−27 The extra energy needed for creating a heterojunction between Cu-BDC and Cu-NDC lattices was obtained by placing six unit cells of Cu-BDC on top of five unit cells of Cu-NDC (see Figure 3). The lattice constants of Cu-NDC and Cu-BDC amount to 1.34 and 1.12 nm, respectively. The corresponding ratio, 1.20 (1.34/ 1.12), is sufficiently close to 5/6 = 0.997 [(1.34*5)/ (1.12*6)]. To warrant the applicability of the 5-on-6 model for the heterojunction, the 5-on-6 matching leaves one undercoordinated (only three instead of four linkers attached) paddle wheel on the Cu-BDC/Cu-NDC interface (see Figure 3 top, dotted blue circle). This vacancy was annihilated by attaching an acetate group. This is a rather plausible way to saturate this unsatisfied valence since the (Cu++)2 dimers are provided in the form of Cu-acetate during the growth process. The presence of such additional acetate groups at the heterojunctions has been shown experimentally using F-labeled acetate groups (Supporting Information Figure S8). Using the geometry indicated in the upper part of Figure 3 as starting point, a geometry optimization was carried out for a model with a thickness of one layer on each side of the heterojunction. The resulting
geometry is shown in Figure 3 (bottom), which demonstrates that the lattice distortion imposed by this rather large lattice mismatch is confined to the immediate vicinity of the junction. There is a significant difference between the interface of the heterojunction between two MOF phases, exemplified for the Cu-BDC-on-Cu-NDC MOF, and typical interfaces between covalently bound materials, as for example semiconductors. In the latter case, unsatisfied valences result in the formation of dangling bonds, leading to a severe energy penalty and also a chemical instability resulting from the pronounced chemical reactivity of dangling bonds. In case of metal−organic frameworks, because of the large pore size enough place is available to annihilate such vacancies by attaching a smaller chemical functionality in this case an acetate group instead of the large organic linkers (H2BDC, H2NDC, or H2BPDC) used to build the MOF framework. As a result, all valences at the MOF/MOF heterointerface can be fully satisfied, and the only contributions to the energy penalty resulting from the lattice mismatch come from excess strain. The strain energy, that is, the summation of the deformation energy of all connectors and linkers compared to the pristine MOF lattice, accounts to a total of 0.68 eV per nm interface length, normalized for one layer. This is significantly less than the energy of a typical covalent bond, which amounts to 3 eV or more. Moreover, the strain is delocalized over a rather large area, allowing for little local stress. The relatively low lattice deformation energy that is necessary to form the heteroepitactic interface can be understood in terms of the very small elastic constants that 1528
dx.doi.org/10.1021/nl404767k | Nano Lett. 2014, 14, 1526−1529
Nano Letters
Letter
(8) Bauer, C. A.; Timofeeva, T. V.; Settersten, T. B.; Patterson, B. D.; Liu, V. H.; Simmons, B. A.; Allendorf, M. D. J. Am. Chem. Soc. 2007, 129 (22), 7136−7144. (9) Tanaka, D.; Henke, A.; Albrecht, K.; Moeller, M.; Nakagawa, K.; Kitagawa, S.; Groll, J. Nature Chem. 2010, 2 (5), 410−416. (10) Yanai, N.; Kitayama, K.; Hijikata, Y.; Sato, H.; Matsuda, R.; Kubota, Y.; Takata, M.; Mizuno, M.; Uemura, T.; Kitagawa, S. Nat. Mater. 2011, 10 (10), 787−793. (11) Devic, T.; Horcajada, P.; Serre, C.; Salles, F.; Maurin, G.; Moulin, B.; Heurtaux, D.; Clet, G.; Vimont, A.; Greneche, J. M.; Le Ouay, B.; Moreau, F.; Magnier, E.; Filinchuk, Y.; Marrot, J.; Lavalley, J. C.; Daturi, M.; Ferey, G. J. Am. Chem. Soc. 2010, 132 (3), 1127−1136. (12) Meilikhov, M.; Furukawa, S.; Hirai, K.; Fischer, R. A.; Kitagawa, S. Angew. Chem. 2013, 125 (1), 359−363. (13) Furukawa, S.; Hirai, K.; Nakagawa, K.; Takashima, Y.; Matsuda, R.; Tsuruoka, T.; Kondo, M.; Haruki, R.; Tanaka, D.; Sakamoto, H.; Shimomura, S.; Sakata, O.; Kitagawa, S. Angew. Chem., Int. Ed. 2009, 48 (10), 1766−1770. (14) Furukawa, S.; Hirai, K.; Takashima, Y.; Nakagawa, K.; Kondo, M.; Tsuruoka, T.; Sakata, O.; Kitagawa, S. Chem. Commun. 2009, No. 34, 5097−5099. (15) Shekhah, O.; Hirai, K.; Wang, H.; Uehara, H.; Kondo, M.; Diring, S.; Zacher, D.; Fischer, R. A.; Sakata, O.; Kitagawa, S.; Furukawa, S.; Wöll, C. Dalton Trans. 2011, 40 (18), 4954−4958. (16) Deng, H. X.; Grunder, S.; Cordova, K. E.; Valente, C.; Furukawa, H.; Hmadeh, M.; Gandara, F.; Whalley, A. C.; Liu, Z.; Asahina, S.; Kazumori, H.; O’Keeffe, M.; Terasaki, O.; Stoddart, J. F.; Yaghi, O. M. Science 2012, 336 (6084), 1018−1023. (17) Kreno, L. E.; Leong, K.; Farha, O. K.; Allendorf, M.; Van Duyne, R. P.; Hupp, J. T. Chem. Rev. 2011, 112 (2), 1105−1125. (18) Tang, S. J.; Lee, C. Y.; Huang, C. C.; Chang, T. R.; Cheng, C. M.; Tsuei, K. D.; Jeng, H. T.; Yeh, V.; Chiang, T. C. Phys. Rev. Lett. 2011, 107 (6), 066802. (19) Shekhah, O.; Wang, H.; Kowarik, S.; Schreiber, F.; Paulus, M.; Tolan, M.; Sternemann, C.; Evers, F.; Zacher, D.; Fischer, R. A.; Wöll, C. J. Am. Chem. Soc. 2007, 129 (49), 15118. (20) Liu, J.; Lukose, B.; Shekhah, O.; Arslan, H. K.; Weidler, P.; Gliemann, H.; Bräse, S.; Grosjean, S.; Godt, A.; Feng, X.; Mullen, K.; Magdau, I.-B.; Heine, T.; Wöll, C. Sci. Rep. 2012, 2, 921. (21) Arslan, H. K.; Shekhah, O.; Wieland, D. C. F.; Paulus, M.; Sternemann, C.; Schroer, M. A.; Tiemeyer, S.; Tolan, M.; Fischer, R. A.; Wöll, C. J. Am. Chem. Soc. 2011, 133 (21), 8158−8161. (22) Arslan, H. K.; Shekhah, O.; Wohlgemuth, J.; Franzreb, M.; Fischer, R. A.; Wöll, C. Adv. Funct. Mater. 2011, 21 (22), 4228. (23) Trent, J. S.; Scheinbeim, J. I.; Couchman, P. R. Macromolecules 1983, 16 (4), 589−598. (24) Adrian, M.; Dubochet, J.; Fuller, S. D.; Harris, J. R. Micron 1998, 29 (2−3), 145−160. (25) Elstner, M.; Porezag, D.; Jungnickel, G.; Elsner, J.; Haugk, M.; Frauenheim, T.; Suhai, S.; Seifert, G. Phys. Rev. B 1998, 58 (11), 7260−7268. (26) Lukose, B.; Supronowicz, B.; St. Petkov, P.; Frenzel, J.; Kuc, A. B.; Seifert, G.; Vayssilov, G. N.; Heine, T. Phys. Status Solidi B 2012, 249 (2), 335−342. (27) Zhechkov, L.; Heine, T.; Patchkovskii, S.; Seifert, G.; Duarte, H. A. J. Chem. Theory Comput. 2005, 1 (5), 841−847. (28) Kuc, A.; Enyashin, A.; Seifert, G. J. Phys. Chem. B 2007, 111 (28), 8179−8186.
are present in MOFs. For SURMOF-2 we obtain, within DFTB values for the bulk modulus of B = 37.4 GPa for SURMOF-2 (BDC) and of B = 22.1 GPa for MOF-2 (NDC). These values are similar to those of other MOFs (B = 34.7, 15.34, 10.10, and 10.73 GPa for HKUST-1, MOF-5, −177, and DUT-6 (MOF205) within DFTB.26 This elastic constant is about 1 order of magnitude lower than those of normal inorganic solids such as silicon or III−V semiconductors.28 In conclusion, we have demonstrated that the idealized, highly crystalline, oriented hierarchically porous MOF crystal structure, Cu-BPDC/Cu-NDC/Cu-BDC and Cu-BDC/CuNDC/Cu-BPDC, with large lattice mismatches of 19.6 and 15.7% for which a synthesis using the conventional route is either difficult or impossible, can be readily prepared by using a liquid phase epitaxy scheme involving a layer-by-layer deposition on a templating organic surface. Such unusually strong lattice mismatches are possible as no chemical bond defect is introduced to the structure, and the associated stress is distributed over a large volume. The availability of MOFcoatings with vertical pore size gradients opens up the possibility to use them in multilevel filtering systems and to reduce or even eliminate cross-talk in MOF-based sensors.8
■
ASSOCIATED CONTENT
S Supporting Information *
Experimental procedures, crystal structure, X-ray diffraction results, SEM images, and details of the theoretical models are included. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS Financial support by Deutsche Forschungsgemeinschaft (DFG) within the Priority Program Metal−Organic Frame-works (SPP 1362) is gratefully acknowledged. Z.W. thanks the Chinese Scholarship Council (CSC) for financial support. T.H. acknowledges funding through the European Research Council (ERC StG C3ENV Grant Agreement 256962). We thank Frank Friedrich, Carlos Azucena, and Engelbert Redel for the SEM measurement and helpful discussion.
■
REFERENCES
(1) Getman, R. B.; Bae, Y. S.; Wilmer, C. E.; Snurr, R. Q. Chem. Rev. 2012, 112 (2), 703−723. (2) Yusenko, K.; Meilikhov, M.; Zacher, D.; Wieland, F.; Sternemann, C.; Stammer, X.; Ladnorg, T.; Wöll, C.; Fischer, R. A. CrystEngComm 2010, 12 (7), 2086−2090. (3) Morris, R. E.; Wheatley, P. S. Angew. Chem., Int. Ed. 2008, 47 (27), 4966−4981. (4) Li, J. R.; Kuppler, R. J.; Zhou, H. C. Chem. Soc. Rev. 2009, 38 (5), 1477−1504. (5) Ferey, G.; Serre, C. Chem. Soc. Rev. 2009, 38 (5), 1380−1399. (6) Allendorf, M. D.; Houk, R. J. T.; Andruszkiewicz, L.; Talin, A. A.; Pikarsky, J.; Choudhury, A.; Gall, K. A.; Hesketh, P. J. J. Am. Chem. Soc. 2008, 130 (44), 14404. (7) Horcajada, P.; Serre, C.; Grosso, D.; Boissiere, C.; Perruchas, S.; Sanchez, C.; Ferey, G. Adv. Mater. 2009, 21 (19), 1931−1935. 1529
dx.doi.org/10.1021/nl404767k | Nano Lett. 2014, 14, 1526−1529
Nanoporous Designer Solids with Huge Lattice Constant Gradients: Multiheteroepitaxy of Metal−Organic Frameworks Zhengbang Wang,† Jinxuan Liu,† Binit Lukose,‡ Zhigang Gu,† Peter G. Weidler,† Hartmut Gliemann,† Thomas Heine,‡ and Christof Wöll*,† †
Karlsruhe Institute of Technology, Institute of Functional Interfaces (IFG), Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany ‡ Center for Functional Nanomaterials, School of Engineering and Science, Jacobs University Bremen, Campus Ring 1, 28759 Bremen, Germany S Supporting Information *
ABSTRACT: We demonstrate the realization of hierarchically organized MOF (metal−organic framework) multilayer systems with pronounced differences in the size of the nanoscale pores. Unusually large values for the lattice constant mismatch at the MOF−MOF heterojunctions are made possible by a particular liquid-phase epitaxy process. The multiheteroepitaxy is demonstrated for the isoreticular SURMOF-2 series [Liu et al. Sci. Rep. 2012, 2, 921] by fabricating trilayer systems with lattice constants of 1.12, 1.34, and 1.55 nm. Despite these large (20%) lattice mismatches, highly crystalline, oriented multilayers were obtained. A thorough theoretical analysis of the MOF-on-MOF heterojunction structure and energetics allows us to identify the two main reasons for this unexpected tolerance of large lattice mismatch: the healing of vacancies with acetate groups and the low elastic constant of MOF materials. KEYWORDS: Metal−organic frameworks, heteroepitaxy, lattice mismatch, thin films, loading ighly porous, crystalline “Designer Solids” with different functionalities fabricated via the assembly of molecular connectors and linkers are receiving a still increasing amount of interest. Going beyond the adsorption of small molecules for gas storage1−3 and separation,4 more advanced applications exploit the enormous flexibility of these metal−organic frameworks, MOFs, for applications in sensorics,5,6 optics,7,8 and energy storage.9−11 The countless number of different MOF types resulting from the combination of metal or metal− oxo connectors with organic linkers can be extended further by employing heteroepitaxy. Growing different types of MOFs in a sequential fashion12−15 allows entering the next level of complexity. Of particular interest here is the stacking of MOFs with different lattice constants. For example in sensorics, it may be required to have a first layer of MOFs with pore sizes sufficiently large to accommodate nanoobjects such as proteins16 or metal−nanoparticles. An additional topping layer with smaller pore sizes could then serve as a filter to reduce cross-talk in MOF-based sensorics.17 The range of lattice constants so far realized in MOFs ranges from less than 1 to 10 nm.16 A combination of MOF materials with different lattice constants is a major challenge, because heteroepitaxy of conventional materials requires lattice constant matching, that is, the lattice constant of the material to be grown must match that of the templating substrate. Particularly important is the case of metal-on-semiconductor epitaxy, where sophisticated strategies are used to overcome the restrictions
H
© 2014 American Chemical Society
imposed by the fact that the second material must match the lattice constant of the first material by better than 2%.18 When the lattice constants differ, the deposited layers are strained, leading to a termination of the growth or low quality of the epitaxially grown layers resulting from high defect densities. Here, we demonstrate that the high elasticity of MOFs strongly relaxes, if not removes, the requirement of lattice constant matching. We show that epitaxial growth can be obtained for lattice constant mismatches as large as 20%. Quantum chemical calculations allow determining the related interface energy penalty. The theoretical results also demonstrate that the intrinsic elasticity of MOFs helps to yield epitaxial growth also in case of strongly different lattice constants. For the heteroepitaxial growth of the hierarchically organized MOF multilayer systems, we use the liquid phase epitaxy (LPE) process, a layer-by-layer procedure that yields oriented, highly crystalline MOF thin films that are referred as SURMOFs.19 Briefly, the SURMOF process consists of spraying appropriately functionalized substrates with solutions of the reactants in a sequential fashion (see Figure S1, Supporting Information). This step-by-step process allows us to separately control the metal and organic linker deposition on the substrate. A huge Received: December 23, 2013 Revised: February 4, 2014 Published: February 10, 2014 1526
dx.doi.org/10.1021/nl404767k | Nano Lett. 2014, 14, 1526−1529
Nano Letters
Letter
advantage of this layer-by-layer process is the possibility to change the composition of the SURMOFs along the vertical direction. In previous work, we have demonstrated the heteroepitaxy of layer-pillar MOFs exhibiting the same lattice constant but different metal ions in the connectors.15 In order to investigate the importance of lattice mismatch in MOF-onMOF heteroepitaxy, here we keep the same connector but used different organic linkers, resulting in changing the lattice constant of the MOF thin films for yielding hierarchically organized MOF multilayer systems. The studies to be presented in this communication have been carried out using the structural platform provided by SURMOF-2, an isoreticular class of metal organic frameworks described in previous work.20 SURMOF-2 is synthesized from Cu-acetate and dicarboxylic acids using liquid phase epitaxy;20 the crystal structures of these compounds are shown in Figure S2 in the Supporting Information. In a first set of experiments, an initial MOF-layer has been grown by LPE from BPDC (biphenyl-4,4′-dicarboxylic acid) and Cu-acetate using the conditions that were published elsewhere.21,22 As evidenced by the out-of-plane (Figure 1, green curve) and in-plane (Supporting Information Figure S3, green curve) X-ray diffractograms, this process yields a crystalline, highly oriented SURMOF-2 with a lattice-constant of 1.55 nm. The out-ofplane XRD data reveals the crystallinity perpendicular to the substrate surface. The fact that only the (001)C and (002)C diffraction peaks are seen, demonstrates the orientated nature of the SURMOF. After 10 deposition cycles, the long BPDC linker was replaced by the shorter NDC (naphthalene-2,6dicarboxylic acid) linker. No change was made to the connector. As evidenced by the X-ray diffractograms (peaks labeled (001)B and (002)B) shown in Figure 1, after another 10 deposition cycles with the shorter NDC ligand a second MOFlayer with a smaller lattice constant of 1.34 nm was formed on the first one. A comparison to the XRD-data recorded for the corresponding SURMOF made directly using Cu-acetate and NDC reveals that the structures are identical. Note, the XRDpeaks located at 5.9° demonstrate that the initial SURMOF layer is still present. In the next step, an even shorter ligand, BDC (benzene-1,4-dicarboxylic acid), was used. Again, the XRD data (peaks labeled (001)A and (002)A) shown in Figure 1 (black curve, out-of plane) and Supporting Information Figure S3 (black curve, in-plane) demonstrate that a third type of MOFs with the same orientation but with a lattice-constant decrease to 1.12 nm has been grown on the supporting substrate (Cu-NDC). The well-defined multilayer structure of the trilayer is directly demonstrated by the SEM micrographs shown in Figure 2. Whereas the cross sections of the hierarchically porous MOF crystal films (Cu-BPDC + Cu-NDC+ Cu-BDC) reveal only little contrast (Figure 2A), selective staining23,24 can be used to make the different porosities of the adlayers visible. To this end, the sample shown in Figure 2A was immersed in a solution of a metal−organic compound, Eu(bzac)3bipy (bzac = 1-benzoylacetone, bipy = 2,2′-bipyridine, for details of staining process see Supporting Information) with a diameter of 0.97 nm. Because of its fairly large size, the Eu-compound is loaded only into the largest pores (Supporting Information Figure S6), and the lower Cu-BPDC-layer (thickness ∼250 nm) can be clearly identified (Figure 2B). The very effective spray-process22 used here yields a thickness of ∼250 nm with only 10 cycles. In order to demonstrate that the constraint of lattice matching is virtually nonexistent for this reticular class of
Figure 1. (Left) Schematic illustration of the construction of hierarchically porous MOF crystal films (Cu-BPDC + Cu-NDC + Cu-BDC) with large lattice mismatches grown on self-assembled monolayer (SAM) modified Au-substrates by using layer-by-layer LPE method (dark blue circles represent metal connectors). Upper right: Molecular structures of the organic ligands. Lower right: Out-of-plane XRD data recorded for the SURMOFs depicted on the left.
MOFs, we have instead of stepwise decreasing the lattice constant as in the example presented above also realized the opposite scenario of increasing lattice constants. Starting with an initial layer of Cu-BDC (with a pore-size of 1.12 nm) we continued to deposit Cu-NDC and finally Cu-BPDC, thus increasing the pore-sizes to 1.34 nm (NDC) and finally 1.55 nm (BPDC). The resulting scenario is shown in Supporting Information Figure S4. As confirmed by the out-of-plane (Figure S4, Supporting Information) and in-plane (Figure S5, Supporting Information) X-ray diffractograms, also for this case of increasing lattice constants the heteroepitaxial growth 1527
dx.doi.org/10.1021/nl404767k | Nano Lett. 2014, 14, 1526−1529
Nano Letters
Letter
Figure 2. (top) Scheme of nanoparticles size-selected loading in hierarchically porous SURMOFs. Down: (A) SEM image recorded for a hierarchically porous SURMOF (Cu-BPDC + Cu-NDC + Cu-BDC) with 5 nm gold films coated on the surface. (B) SEM image recorded for a hierarchically porous SURMOF (Cu-BPDC + Cu-NDC + CuBDC) after loading Eu(bzac)3bipy (bzac = 1-benzoylacetone, bipy = 2,2′-bipyridine) compound without coating gold films on the surface.
Figure 3. (top) Schematic illustration of the formation of flexible lattice matching junction in MOF-on-MOF epitaxy. (bottom) Molecular arrangement as obtained from the quantum chemical calculations (see text) for the heterojunction between Cu-BDC and Cu-NDC.
proceeds without difficulties despite the rather pronounced lattice mismatch. To better understand why for the MOF-on-MOF heteroepitaxy the lattice matching requirement appears to be less stringent than for other cases of heteroepitaxy, we have carried out an extensive set of quantum chemical calculations for the heterojunction between Cu-BDC and Cu-NDC. The method used here is an approximate variant of density functional theory (DFT), London dispersion-corrected self-consistent charge density-functional based tight-binding (DFTB).25−27 The extra energy needed for creating a heterojunction between Cu-BDC and Cu-NDC lattices was obtained by placing six unit cells of Cu-BDC on top of five unit cells of Cu-NDC (see Figure 3). The lattice constants of Cu-NDC and Cu-BDC amount to 1.34 and 1.12 nm, respectively. The corresponding ratio, 1.20 (1.34/ 1.12), is sufficiently close to 5/6 = 0.997 [(1.34*5)/ (1.12*6)]. To warrant the applicability of the 5-on-6 model for the heterojunction, the 5-on-6 matching leaves one undercoordinated (only three instead of four linkers attached) paddle wheel on the Cu-BDC/Cu-NDC interface (see Figure 3 top, dotted blue circle). This vacancy was annihilated by attaching an acetate group. This is a rather plausible way to saturate this unsatisfied valence since the (Cu++)2 dimers are provided in the form of Cu-acetate during the growth process. The presence of such additional acetate groups at the heterojunctions has been shown experimentally using F-labeled acetate groups (Supporting Information Figure S8). Using the geometry indicated in the upper part of Figure 3 as starting point, a geometry optimization was carried out for a model with a thickness of one layer on each side of the heterojunction. The resulting
geometry is shown in Figure 3 (bottom), which demonstrates that the lattice distortion imposed by this rather large lattice mismatch is confined to the immediate vicinity of the junction. There is a significant difference between the interface of the heterojunction between two MOF phases, exemplified for the Cu-BDC-on-Cu-NDC MOF, and typical interfaces between covalently bound materials, as for example semiconductors. In the latter case, unsatisfied valences result in the formation of dangling bonds, leading to a severe energy penalty and also a chemical instability resulting from the pronounced chemical reactivity of dangling bonds. In case of metal−organic frameworks, because of the large pore size enough place is available to annihilate such vacancies by attaching a smaller chemical functionality in this case an acetate group instead of the large organic linkers (H2BDC, H2NDC, or H2BPDC) used to build the MOF framework. As a result, all valences at the MOF/MOF heterointerface can be fully satisfied, and the only contributions to the energy penalty resulting from the lattice mismatch come from excess strain. The strain energy, that is, the summation of the deformation energy of all connectors and linkers compared to the pristine MOF lattice, accounts to a total of 0.68 eV per nm interface length, normalized for one layer. This is significantly less than the energy of a typical covalent bond, which amounts to 3 eV or more. Moreover, the strain is delocalized over a rather large area, allowing for little local stress. The relatively low lattice deformation energy that is necessary to form the heteroepitactic interface can be understood in terms of the very small elastic constants that 1528
dx.doi.org/10.1021/nl404767k | Nano Lett. 2014, 14, 1526−1529
Nano Letters
Letter
(8) Bauer, C. A.; Timofeeva, T. V.; Settersten, T. B.; Patterson, B. D.; Liu, V. H.; Simmons, B. A.; Allendorf, M. D. J. Am. Chem. Soc. 2007, 129 (22), 7136−7144. (9) Tanaka, D.; Henke, A.; Albrecht, K.; Moeller, M.; Nakagawa, K.; Kitagawa, S.; Groll, J. Nature Chem. 2010, 2 (5), 410−416. (10) Yanai, N.; Kitayama, K.; Hijikata, Y.; Sato, H.; Matsuda, R.; Kubota, Y.; Takata, M.; Mizuno, M.; Uemura, T.; Kitagawa, S. Nat. Mater. 2011, 10 (10), 787−793. (11) Devic, T.; Horcajada, P.; Serre, C.; Salles, F.; Maurin, G.; Moulin, B.; Heurtaux, D.; Clet, G.; Vimont, A.; Greneche, J. M.; Le Ouay, B.; Moreau, F.; Magnier, E.; Filinchuk, Y.; Marrot, J.; Lavalley, J. C.; Daturi, M.; Ferey, G. J. Am. Chem. Soc. 2010, 132 (3), 1127−1136. (12) Meilikhov, M.; Furukawa, S.; Hirai, K.; Fischer, R. A.; Kitagawa, S. Angew. Chem. 2013, 125 (1), 359−363. (13) Furukawa, S.; Hirai, K.; Nakagawa, K.; Takashima, Y.; Matsuda, R.; Tsuruoka, T.; Kondo, M.; Haruki, R.; Tanaka, D.; Sakamoto, H.; Shimomura, S.; Sakata, O.; Kitagawa, S. Angew. Chem., Int. Ed. 2009, 48 (10), 1766−1770. (14) Furukawa, S.; Hirai, K.; Takashima, Y.; Nakagawa, K.; Kondo, M.; Tsuruoka, T.; Sakata, O.; Kitagawa, S. Chem. Commun. 2009, No. 34, 5097−5099. (15) Shekhah, O.; Hirai, K.; Wang, H.; Uehara, H.; Kondo, M.; Diring, S.; Zacher, D.; Fischer, R. A.; Sakata, O.; Kitagawa, S.; Furukawa, S.; Wöll, C. Dalton Trans. 2011, 40 (18), 4954−4958. (16) Deng, H. X.; Grunder, S.; Cordova, K. E.; Valente, C.; Furukawa, H.; Hmadeh, M.; Gandara, F.; Whalley, A. C.; Liu, Z.; Asahina, S.; Kazumori, H.; O’Keeffe, M.; Terasaki, O.; Stoddart, J. F.; Yaghi, O. M. Science 2012, 336 (6084), 1018−1023. (17) Kreno, L. E.; Leong, K.; Farha, O. K.; Allendorf, M.; Van Duyne, R. P.; Hupp, J. T. Chem. Rev. 2011, 112 (2), 1105−1125. (18) Tang, S. J.; Lee, C. Y.; Huang, C. C.; Chang, T. R.; Cheng, C. M.; Tsuei, K. D.; Jeng, H. T.; Yeh, V.; Chiang, T. C. Phys. Rev. Lett. 2011, 107 (6), 066802. (19) Shekhah, O.; Wang, H.; Kowarik, S.; Schreiber, F.; Paulus, M.; Tolan, M.; Sternemann, C.; Evers, F.; Zacher, D.; Fischer, R. A.; Wöll, C. J. Am. Chem. Soc. 2007, 129 (49), 15118. (20) Liu, J.; Lukose, B.; Shekhah, O.; Arslan, H. K.; Weidler, P.; Gliemann, H.; Bräse, S.; Grosjean, S.; Godt, A.; Feng, X.; Mullen, K.; Magdau, I.-B.; Heine, T.; Wöll, C. Sci. Rep. 2012, 2, 921. (21) Arslan, H. K.; Shekhah, O.; Wieland, D. C. F.; Paulus, M.; Sternemann, C.; Schroer, M. A.; Tiemeyer, S.; Tolan, M.; Fischer, R. A.; Wöll, C. J. Am. Chem. Soc. 2011, 133 (21), 8158−8161. (22) Arslan, H. K.; Shekhah, O.; Wohlgemuth, J.; Franzreb, M.; Fischer, R. A.; Wöll, C. Adv. Funct. Mater. 2011, 21 (22), 4228. (23) Trent, J. S.; Scheinbeim, J. I.; Couchman, P. R. Macromolecules 1983, 16 (4), 589−598. (24) Adrian, M.; Dubochet, J.; Fuller, S. D.; Harris, J. R. Micron 1998, 29 (2−3), 145−160. (25) Elstner, M.; Porezag, D.; Jungnickel, G.; Elsner, J.; Haugk, M.; Frauenheim, T.; Suhai, S.; Seifert, G. Phys. Rev. B 1998, 58 (11), 7260−7268. (26) Lukose, B.; Supronowicz, B.; St. Petkov, P.; Frenzel, J.; Kuc, A. B.; Seifert, G.; Vayssilov, G. N.; Heine, T. Phys. Status Solidi B 2012, 249 (2), 335−342. (27) Zhechkov, L.; Heine, T.; Patchkovskii, S.; Seifert, G.; Duarte, H. A. J. Chem. Theory Comput. 2005, 1 (5), 841−847. (28) Kuc, A.; Enyashin, A.; Seifert, G. J. Phys. Chem. B 2007, 111 (28), 8179−8186.
are present in MOFs. For SURMOF-2 we obtain, within DFTB values for the bulk modulus of B = 37.4 GPa for SURMOF-2 (BDC) and of B = 22.1 GPa for MOF-2 (NDC). These values are similar to those of other MOFs (B = 34.7, 15.34, 10.10, and 10.73 GPa for HKUST-1, MOF-5, −177, and DUT-6 (MOF205) within DFTB.26 This elastic constant is about 1 order of magnitude lower than those of normal inorganic solids such as silicon or III−V semiconductors.28 In conclusion, we have demonstrated that the idealized, highly crystalline, oriented hierarchically porous MOF crystal structure, Cu-BPDC/Cu-NDC/Cu-BDC and Cu-BDC/CuNDC/Cu-BPDC, with large lattice mismatches of 19.6 and 15.7% for which a synthesis using the conventional route is either difficult or impossible, can be readily prepared by using a liquid phase epitaxy scheme involving a layer-by-layer deposition on a templating organic surface. Such unusually strong lattice mismatches are possible as no chemical bond defect is introduced to the structure, and the associated stress is distributed over a large volume. The availability of MOFcoatings with vertical pore size gradients opens up the possibility to use them in multilevel filtering systems and to reduce or even eliminate cross-talk in MOF-based sensors.8
■
ASSOCIATED CONTENT
S Supporting Information *
Experimental procedures, crystal structure, X-ray diffraction results, SEM images, and details of the theoretical models are included. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS Financial support by Deutsche Forschungsgemeinschaft (DFG) within the Priority Program Metal−Organic Frame-works (SPP 1362) is gratefully acknowledged. Z.W. thanks the Chinese Scholarship Council (CSC) for financial support. T.H. acknowledges funding through the European Research Council (ERC StG C3ENV Grant Agreement 256962). We thank Frank Friedrich, Carlos Azucena, and Engelbert Redel for the SEM measurement and helpful discussion.
■
REFERENCES
(1) Getman, R. B.; Bae, Y. S.; Wilmer, C. E.; Snurr, R. Q. Chem. Rev. 2012, 112 (2), 703−723. (2) Yusenko, K.; Meilikhov, M.; Zacher, D.; Wieland, F.; Sternemann, C.; Stammer, X.; Ladnorg, T.; Wöll, C.; Fischer, R. A. CrystEngComm 2010, 12 (7), 2086−2090. (3) Morris, R. E.; Wheatley, P. S. Angew. Chem., Int. Ed. 2008, 47 (27), 4966−4981. (4) Li, J. R.; Kuppler, R. J.; Zhou, H. C. Chem. Soc. Rev. 2009, 38 (5), 1477−1504. (5) Ferey, G.; Serre, C. Chem. Soc. Rev. 2009, 38 (5), 1380−1399. (6) Allendorf, M. D.; Houk, R. J. T.; Andruszkiewicz, L.; Talin, A. A.; Pikarsky, J.; Choudhury, A.; Gall, K. A.; Hesketh, P. J. J. Am. Chem. Soc. 2008, 130 (44), 14404. (7) Horcajada, P.; Serre, C.; Grosso, D.; Boissiere, C.; Perruchas, S.; Sanchez, C.; Ferey, G. Adv. Mater. 2009, 21 (19), 1931−1935. 1529
dx.doi.org/10.1021/nl404767k | Nano Lett. 2014, 14, 1526−1529
